Multilevel Index Algorithm Based on Improved Association Rule Mining

Abstract

Association rule mining is an important research content in data analysis. At the same time it is also the key technology of database index, proposed an improved association rule mining algorithm. Using multi index database frequent item sets of association rules produce the desired set of features, the feature sub space narrowing of candidate item sets in scale compression method. Based on the repeated scanning of index database, index node according to the minimum support threshold filtering rules and constraints, depth first traversal of the index database frequent item list, according to the data distribution structure rearrangement association rules list, improve the efficiency of data index. Simulation results show that using the method of text information database index, it has better accuracy in the mining of association rules. It can improve the retrieval precision and recall, and the construction of the database access. It has very good application value.